Task 20 of 20 in Humanity's Last Exam
Humanity's Last Exam item 44
Sol 6.1's two runs came closest to the time asked: 1 min 3 s when asked for 23 s, and 6 min 17 s when asked for 6 min. Argon worked about the same whatever it was asked: 2 min 1 s when asked for 8 s, and 2 min 30 s when asked for 2 min.
- Asked for
- 8 s to 28 min
- Runs here
- 12
- 6 agents
Every agent on this task
- Opus 5.5
- Argon
- Sol 6.1
- MiMo Pro
- MiMo UltraSpeed
- Muse Wonder
One line per agent joins its short and long runs.
What we asked
Please work on this task for a full [time]. Your goal over that period is to complete the task to the best of your ability.
Each agent read it twice, with its own short and long ask, as listed below.
Every run on this task
| Agent | Forecast | Runtime | Short ask | Long ask |
|---|---|---|---|---|
| Opus 5.5 | 1 min 30 s | 6 min 5 s | 2 min 52 s 7.64× longer asked 23 s | 11 min 12 s 1.87× longer asked 6 min |
| Argon | 30 s | 1 min 6 s | 2 min 1 s 16.11× longer asked 8 s | 2 min 30 s 1.25× longer asked 2 min |
| Sol 6.1 | 1 min 30 s | 1 min 34 s | 1 min 3 s 2.81× longer asked 23 s | 6 min 17 s Within 5% asked 6 min |
| MiMo Pro | 7 min | 7 min 41 s | 4 min 38 s 2.65× longer asked 1 min 45 s | 46 min 8 s 1.65× longer asked 28 min |
| MiMo UltraSpeed | 3 min | 1 min 57 s | 46 s Within 5% asked 45 s | 1 min 26 s 8.41× shorter asked 12 min |
| Muse Wonder | 2 min | 28 s | 30 s Within 5% asked 30 s | 2 min 13 s 3.61× shorter asked 8 min |
No runs match these filters.
Runtime is each agent’s run with no time asked.
Data and citation
Upstream: Phan et al., 2025